Ultra-high speed Monte Carlo computing techniques using field programmable gate arrays

Alexander Pasciak, John R. Ford

Research output: Contribution to conferencePaper

Abstract

Advancements in parallel and cluster computing have made many complex Monte Carlo simulations possible in the past several years. Unfortunately, cluster computers are large, expensive, and still not fast enough to make Monte Carlo useful for calculations requiring a near real time evaluation period. For Monte Carlo simulations, a small computational unit called a Field Programmable Gate Array (FPGA) is capable of bringing the power of a large cluster computer into any desktop PC. Because an FPGA is capable of executing Monte Carlo simulations with a high degree of parallelism, a simulation run on a large FPGA can be executed at a much higher rate than an equivalent simulation on a modern single processor desktop PC. In this paper we discuss a simple radiation transport problem involving moderate energy photons incident on a 3 dimensional target. By comparing the evaluation speed of this transport problem on an FPGA to the evaluation speed of the same transport problem using standard computing techniques, we show that it is possible to accelerate Monte Carlo computations significantly using field programmable gate arrays. In fact, we have found that our simple photon transport test case can be evaluated about 650 times faster on a large FPGA than it can on a 3.2 Ghz Pentium-4 desktop PC running MCNP5-an acceleration factor that we predict will be largely preserved for many Monte Carlo simulations.

Original languageEnglish (US)
Pages1063-1073
Number of pages11
StatePublished - Dec 1 2005
Externally publishedYes
EventMonte Carlo 2005 Topical Meeting - Chattanooga, TN, United States
Duration: Apr 17 2005Apr 21 2005

Other

OtherMonte Carlo 2005 Topical Meeting
CountryUnited States
CityChattanooga, TN
Period4/17/054/21/05

Fingerprint

Field programmable gate arrays (FPGA)
Photons
Cluster computing
Parallel processing systems
Radiation
Monte Carlo simulation

All Science Journal Classification (ASJC) codes

  • Engineering(all)

Cite this

Pasciak, A., & Ford, J. R. (2005). Ultra-high speed Monte Carlo computing techniques using field programmable gate arrays. 1063-1073. Paper presented at Monte Carlo 2005 Topical Meeting, Chattanooga, TN, United States.

Ultra-high speed Monte Carlo computing techniques using field programmable gate arrays. / Pasciak, Alexander; Ford, John R.

2005. 1063-1073 Paper presented at Monte Carlo 2005 Topical Meeting, Chattanooga, TN, United States.

Research output: Contribution to conferencePaper

Pasciak, A & Ford, JR 2005, 'Ultra-high speed Monte Carlo computing techniques using field programmable gate arrays' Paper presented at Monte Carlo 2005 Topical Meeting, Chattanooga, TN, United States, 4/17/05 - 4/21/05, pp. 1063-1073.
Pasciak A, Ford JR. Ultra-high speed Monte Carlo computing techniques using field programmable gate arrays. 2005. Paper presented at Monte Carlo 2005 Topical Meeting, Chattanooga, TN, United States.
Pasciak, Alexander ; Ford, John R. / Ultra-high speed Monte Carlo computing techniques using field programmable gate arrays. Paper presented at Monte Carlo 2005 Topical Meeting, Chattanooga, TN, United States.11 p.
@conference{74126f072a404e51b038021ac63331c1,
title = "Ultra-high speed Monte Carlo computing techniques using field programmable gate arrays",
abstract = "Advancements in parallel and cluster computing have made many complex Monte Carlo simulations possible in the past several years. Unfortunately, cluster computers are large, expensive, and still not fast enough to make Monte Carlo useful for calculations requiring a near real time evaluation period. For Monte Carlo simulations, a small computational unit called a Field Programmable Gate Array (FPGA) is capable of bringing the power of a large cluster computer into any desktop PC. Because an FPGA is capable of executing Monte Carlo simulations with a high degree of parallelism, a simulation run on a large FPGA can be executed at a much higher rate than an equivalent simulation on a modern single processor desktop PC. In this paper we discuss a simple radiation transport problem involving moderate energy photons incident on a 3 dimensional target. By comparing the evaluation speed of this transport problem on an FPGA to the evaluation speed of the same transport problem using standard computing techniques, we show that it is possible to accelerate Monte Carlo computations significantly using field programmable gate arrays. In fact, we have found that our simple photon transport test case can be evaluated about 650 times faster on a large FPGA than it can on a 3.2 Ghz Pentium-4 desktop PC running MCNP5-an acceleration factor that we predict will be largely preserved for many Monte Carlo simulations.",
author = "Alexander Pasciak and Ford, {John R.}",
year = "2005",
month = "12",
day = "1",
language = "English (US)",
pages = "1063--1073",
note = "Monte Carlo 2005 Topical Meeting ; Conference date: 17-04-2005 Through 21-04-2005",

}

TY - CONF

T1 - Ultra-high speed Monte Carlo computing techniques using field programmable gate arrays

AU - Pasciak, Alexander

AU - Ford, John R.

PY - 2005/12/1

Y1 - 2005/12/1

N2 - Advancements in parallel and cluster computing have made many complex Monte Carlo simulations possible in the past several years. Unfortunately, cluster computers are large, expensive, and still not fast enough to make Monte Carlo useful for calculations requiring a near real time evaluation period. For Monte Carlo simulations, a small computational unit called a Field Programmable Gate Array (FPGA) is capable of bringing the power of a large cluster computer into any desktop PC. Because an FPGA is capable of executing Monte Carlo simulations with a high degree of parallelism, a simulation run on a large FPGA can be executed at a much higher rate than an equivalent simulation on a modern single processor desktop PC. In this paper we discuss a simple radiation transport problem involving moderate energy photons incident on a 3 dimensional target. By comparing the evaluation speed of this transport problem on an FPGA to the evaluation speed of the same transport problem using standard computing techniques, we show that it is possible to accelerate Monte Carlo computations significantly using field programmable gate arrays. In fact, we have found that our simple photon transport test case can be evaluated about 650 times faster on a large FPGA than it can on a 3.2 Ghz Pentium-4 desktop PC running MCNP5-an acceleration factor that we predict will be largely preserved for many Monte Carlo simulations.

AB - Advancements in parallel and cluster computing have made many complex Monte Carlo simulations possible in the past several years. Unfortunately, cluster computers are large, expensive, and still not fast enough to make Monte Carlo useful for calculations requiring a near real time evaluation period. For Monte Carlo simulations, a small computational unit called a Field Programmable Gate Array (FPGA) is capable of bringing the power of a large cluster computer into any desktop PC. Because an FPGA is capable of executing Monte Carlo simulations with a high degree of parallelism, a simulation run on a large FPGA can be executed at a much higher rate than an equivalent simulation on a modern single processor desktop PC. In this paper we discuss a simple radiation transport problem involving moderate energy photons incident on a 3 dimensional target. By comparing the evaluation speed of this transport problem on an FPGA to the evaluation speed of the same transport problem using standard computing techniques, we show that it is possible to accelerate Monte Carlo computations significantly using field programmable gate arrays. In fact, we have found that our simple photon transport test case can be evaluated about 650 times faster on a large FPGA than it can on a 3.2 Ghz Pentium-4 desktop PC running MCNP5-an acceleration factor that we predict will be largely preserved for many Monte Carlo simulations.

UR - http://www.scopus.com/inward/record.url?scp=33644592076&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33644592076&partnerID=8YFLogxK

M3 - Paper

SP - 1063

EP - 1073

ER -